Neural Architecture Search for Deep Image Prior

By Kary Ho, Andrew Gilbert, Hailin Jin, John Philip Collomosse. 2020

43 citations recorded by the source corpus.

We present a neural architecture search (NAS) technique to enhance the performance of unsupervised image de-noising, in-painting and super-resolution under the recently proposed Deep Image Prior (DIP). We show that evolutionary search can automatically optimize the encoder-decoder (E-D) structure and meta-parameters of the DIP network, which serves as a content-specific prior to regularize these single image restoration tasks. Our binary representation encodes the design space for an asymmetric E-D network that typically converges to yield a content-specific DIP within 10-20 generations using a population size of 500. The optimized architectures consistently improve upon the visual quality of classical DIP for a diverse range of photographic and artistic content.

Subfields: Computer Vision and Pattern Recognition, Model Architectures, Neural Architecture Search & AutoML

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

1. Yujiu Yang
2. Ming-Hsuan Yang
3. Weihao Xia
4. Yulun Zhang
5. Jing-Hao Xue
6. Luc Van Gool
7. Bolei Zhou
8. Lei Zhang

In Model Architectures

1. Bolei Zhou
2. Xingquan Zhu
3. Ming-Hsuan Yang
4. Jing-Hao Xue
5. Yulun Zhang
6. Yujiu Yang
7. Weihao Xia

Citations in the corpus

• GAN Inversion: A Survey (2021, methodology): “[204] automatically optimize the encoder-decoder structure using evolutionary search and meta-parameters of the DIP network, which serves as a content-specific prior to regularize image restoration.”
• A Survey on Evolutionary Neural Architecture Search (2020, unknown): “objective EA GAs [9], [19]–[23], [33]–[35], [47], [49], [52], [56], [59]–[61], [63], [65]–[67], [69]–[71], [74], [78], [81], [82], [85], [86], [88], [90], [92], [93], [95], [97]– [99], [103], [107], [110], [113], [117]– [119], [121], [122], [125]–[127], [131], [138], [143], [144] [39], [145] [44], [”
• Evolutionary Architecture Search for Graph Neural Networks (2020, methodology): “EA-based NAS has been widely used for identifying suitable CNN model for a specific task such as image denoising, in-painting and super-resolution [35].”